TH1E PROCESS OF URBANIZATION: A COMPARATIVE STUDY WITH SPECIAL REFERENCE TO THE UNITED STATES, JAPAN AND INDIA: 1890-1950 by MARK LAWNER B.A., Harvard University, 1960 SUBMITTED IN PARTIAL FULFILLAENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF CITY PLANNING at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY June, 1963 Signature of Author.......... ................ ........ Department of City Planning, May 17, 1963 Certified by..... .; . ; .................. Thesis Supervisor Accepted by......... ....................................... Cha rman, Departmental Committee on Theses -Dedication- To Harry, Irene, Lynne and Robert i Table of Contents Page Introduction....... .......................... 1 Review of the literature.................*.. 3 Propositions and methodology for their testing.............................. 22 Findings..................................... 28 Conclusion... 6 .............. ............... 41 Footnotes Bibliography ..... there is no such thing as "proper territorial limits" of a community; but only approximate limits which are constantly shifting, A national population is continuous in its spread over its national terri- tory, varying in density in what we call "communities"; these "communities" are not discrete entities, but make sense only in reference to the rest of the coun- try. Yet we have found a useful thing to measure in these "communities" and by first considering them as discontinuous entities, we shall perhaps later be better able to comprehend them as they appear in the continuous cobweb of interrelated human organization. -George K. Zipf Acknowledgement The author wishes to thank his advisor, Professor John Friedmann, for reading and reviewing parts of the manuscript throughout the various stages of their completion. The author would also like to express his appreciation to Professors Aaron Fleisher and George Wadsworth for their review of the statistical aspects of the study. For the special application of the Lorenz curves in the text the author is indebted to Jim Watanabe of the Harvard School of Design. Finally, thanks are due to past and present classmates, Ross Harris, Bill Pokross and Paul Kolp, who have made suggestions about, and corrections of, the manuscript. -Abstract- The process of urbanization is a dynamic which is reflected in certain demographic outcomes. Therefore, a definition of this process is best couched in demographic terms. A few key vari- ables and their permutations suffice to describe the process of urbanization for purposes of inter- national comparison--total population, number of cities, city sizes, and city rates of growth. Six propositions derive from these variables, and they have been tested in the United States, Japan and India over the six decades between 1888 and 1950. Stewart's rule relating number of cities to the urban total population ratio is confirmed in all three countries. Tisdale's preconditions-- increase in cities and in city size--are to be found in the United States and Japan but not in India except for the last two decades. While Zipf's rank- size rule does not accurately predict the city size distributions for the most part, considerable rank- order stability prevails among the individual cities in all three countries. Jefferson's "law of the primate city" predicts relative position in the United States and in India (with "dual primacy" as a quali- fication), but "primacy'does not increase steadily with time in any of the three countries. Finally Madden's generalization of retarding growth rates is confined to the U.S. experience; individual rates of growth have yet to peak in Japan and India. A description of the process of urbanization in the above terms is a necessary first step in the deline- ation of the performance characteristics of a "system of cities." -1- -Introduction- "The problem of political economy is posed not by the necessity of an individual making a choice under a given system of constraints but rather by the necessity of a society of people making a choice among alternative systems of constraints. For a legislative body to evaluate an existing system of constraints is for the members of that body to compare....the working properties of this system with the working pro- perties of some altered form of the system. A legislative program is literally a system of constraints upon and prescriptions of the actions of individuals. The outcome of the adoption of one system rather than another is described in terms of limiting forms of distributions. 1 " The system selected for study here encompasses the cities of a nation-state; the performance characteristics of these cities will be analyzed both in the aggregate and by selective size-class. The concept of a "system of cities" has been advanced by many writers, but the structure of such a system, not to mention its universality, is still in the process of being determined. We shall focus in this study on certain strategic demographic variables, e.g. number of cities, city sizes and decade rates of growth, total population,and the variety of ways in which they interrelate. ..... whatever else may be entailed, there is general agreement that the factors which under- lie any such system (of cities) are probably reflected in one way or another in the demo- graphic...relations among cities. It is this general hypothesis that is the primary rationale for studying the relations. 2 -2- The above variables subsume numerous other demographic and economic variables such as natural increase, net migration and numbers and types of economic functions. Also they are considered independently of various other economic variables with which they may or may not correlate. The strategy employed here is to interrelate the number of cities, their population sizes and decade rates of growth as a first step in the delineation of a "system of cities". These demo- graphic variables clearly do not exist in isolation; it is likely that they both partially reflect and condition functional relations among cities. But this aspect would take us well beyond the scope of this study. Consequently, the demographic variables will not be explicitly related to other, "outside" variables, but some use will be made of economic, historical and cultural factors in the inter- pretation of the study findings. Various combinations of the four demographic variables will be shown below to define the process of urbanization, and the case upon which this definition rests will be put forward. The process of urbanization involves, above all, change over time, yet in the midst of this change, certain relationships persist. It is these parameters that we want to determine as well as the correlations of individual variables with each other; that is, this study will attempt to ascertain the underlying parameters of the process of urbanization in addition to establishing various linear correlations. Most of the empirical relation- ships to be tested in this study have been developed by writers who have generalized from the U.S. experience of urbanization. One of the main objects of the study will be to see if, and to what degree, these gener- alizations are applicable to India and Japan. These two countries differ markedly from the U.S. in a number of important respects: among them land area, total population, degree of industrialization, cultural role of cities, historical development in general, etc. In view of such profound differences, it remains to be seen whether or not their urbanization has been similar to that of the U.S. Except for the work of Kingsley Davis and his associates, compar- ative studies of urbanization are conspicuously absent from the literature. This is unfortunate, because the comparative approach allows one to study urbanization within a broad context and thus cull out the universal from the particular. The review of the existing literature will bring out the fact that research to date has focussed on problems of definition, the application of a single concept to one or more countries or the application of several concepts to a single country, but not the testing of a series of related or complementary propositions to several countries. The ideal, of course, would be to test a long series of interrelated propositions with reference to all the countries in the world. This study represents a modest beginning in that direction. -Review of the Literature- Hope Tisdale's definition of urbanization will serve as an entree into the literature, inasmuch as it is often cited by later writers on the subject; some build upon it, others try to substitute alternative formulations. ~04- Urbanization is a process of population concen- tration. It proceeds in two ways: the multipli- cation of points of concentration and the increase in size of individual concentrations. Consistent with the definition of urbanization, cities may be defined as points of concentration. There is no need at this juncture to fix lower limits to the size and density which qualify a concentration as a city. There is no clear cut level of concentration at which a city suddenly springs into being. It is convenient from time to time to name certain levels beyond which con- centrations are designated as cities. This is necessary in analyzing data and identifying charac- teristics of various size groups, but it does not alter the validity of the original concept.3 She defends this kind of demographic definition on grounds that it a) avoids postulating the pro-existence of cities; b) disentangles causes and effects from urbanization itself; c) dif- ferentiates the end product, cities, from the process, urbanization; and d) avoids prejudging certain relationships such as the economic correlates of city size. Her definition has the added advantage of subsuming intervening variables such as natural increase and net migration. To index an increase in the number and size of individual concentrations, Madden makes use of the mean size of all cities over 5,000 persons and also a modified Lorenz curve with percentage of urban population plotted on the vertical scale and percentage of urban places over 5,000 plotted on the horizontal scale.